Extracellular vesicles released by polycyclic aromatic hydrocarbons-treated hepatocytes trigger oxidative stress in recipient hepatocytes by delivering iron

HAL Id: hal-02930146

https://hal.archives-ouvertes.fr/hal-02930146

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Extracellular vesicles released by polycyclic aromatic

hydrocarbons-treated hepatocytes trigger oxidative

stress in recipient hepatocytes by delivering iron

Nettie van Meteren, Dominique Lagadic-Gossmann, Normand Podechard,

Dimitri Gobart, Isabelle Gallais, Martine Chevanne, Aurore Collin, Agnès

Burel, Aurélien Dupont, Ludivine Rault, et al.

To cite this version:

Nettie van Meteren, Dominique Lagadic-Gossmann, Normand Podechard, Dimitri Gobart, Isabelle Gallais, et al.. Extracellular vesicles released by polycyclic aromatic hydrocarbons-treated hepatocytes trigger oxidative stress in recipient hepatocytes by delivering iron. Free Radical Biology and Medicine, Elsevier, 2020, 160, pp.246-262. �10.1016/j.freeradbiomed.2020.08.001�. �hal-02930146�

1

Extracellular vesicles released by polycyclic aromatic hydrocarbons-treated hepatocytes trigger oxidative stress in recipient hepatocytes by delivering iron

Nettie van Meterena, Dominique Lagadic-Gossmanna, Normand Podecharda, Dimitri Gobarta, Isabelle Gallaisa Martine Chevannea, Aurore Collina,1, Agnès Burelb, Aurélien Dupontb, Ludivine Raultd, Soizic Chevancec, Fabienne Gauffrec, Eric Le Ferreca, Odile Sergenta.

a _{Univ Rennes, Inserm, EHESP, Irset (Institut de recherche en santé, environnement et travail)}

- UMR_S 1085, F-35000 Rennes, France

b _{Univ Rennes, Biosit – UMS 3480, US_S 018, F-35000 Rennes, France}

c _{Univ Rennes, CNRS, ISCR (Institut des sciences chimiques de Rennes) – UMR 6226, F-35000}

Rennes, France

d _{Univ Rennes, ScanMAT – UMS 2001, F-35000-Rennes, France}

1 _{Present address : Université Clermont-Auvergne, NEURO-DOL, Inserm U1107, 2, rue Braga,}

63100 Clermont-Ferrand, France.

Corresponding author : Odile Sergent ; IRSET, Faculté de Pharmacie ; 2, av Pr. Léon

Bernard; 35043 Rennes Cedex ; France

E- mail address : odile.sergent@univ-rennes1.fr

Telephone : 33(0)223234808 Fax : 33(0)223235055

Accepted

2 ABSTRACT

A growing body of evidences indicate the major role of extracellular vesicles (EVs) as players

of cell communication in the pathogenesis of liver diseases. EVs are membrane-enclosed

vesicles released by cells into the extracellular environment. Oxidative stress is also a key

component of liver disease pathogenesis, but no role for hepatocyte-derived EVs has yet been

described in the development of this process. Recently, some polycyclic aromatic hydrocarbons

(PAHs), widespread environmental contaminants, were demonstrated to induce EV release

from hepatocytes. They are also well-known to trigger oxidative stress leading to cell death.

Therefore, the aim of this work was to investigate the involvement of EVs derived from

PAHs-treated hepatocytes (PAH-EVs) in possible oxidative damages of healthy recipient hepatocytes,

using both WIF-B9 and primary rat hepatocytes. We first showed that the release of EVs from

PAHs -treated hepatocytes depended on oxidative stress. PAH-EVs were enriched in proteins

related to oxidative stress such as NADPH oxidase and ferritin. They were also demonstrated

to contain more iron. PAH-EVs could then induce oxidative stress in recipient hepatocytes,

thereby leading to apoptosis. Mitochondria and lysosomes of recipient hepatocytes exhibited

significant structural alterations. All those damages were dependent on internalization of EVs

that reached lysosomes with their cargoes. Lysosomes thus appeared as critical organelles for

EVs to induce apoptosis. In addition, pro-oxidant components of PAH-EVs, e.g. NADPH

oxidase and iron, were revealed to be necessary for this cell death.

Key words : Polycyclic aromatic hydrocarbons, extracellular vesicles, hepatocytes, NADPH oxidase, ferritin, LMW iron, lipid peroxidation, lysosomes, mitochondria

Abbreviations : AhR , aryl hydrocarbon receptor; APO, apocynin; BAF, bafilomycin A1; BP, benzo(a)pyrene; CyB, cytochalasin B; CYP1, cytochrome P450 family 1; DBA, dibenzo(a,h)anthracene; DFO, desferrioxamine; DFP, deferiprone; EDS, energy dispersive X-ray

Accepted

3

spectroscopy; EROD, ethoxyresorufin-O-deethylase; EV, extracellular vesicle; FTL, ferritin light chain; LMW iron, low-molecular-weight iron; MCD, methyl-β-cyclodextrin; NTA, nanoparticle tracking analysis; PAH, polycyclic aromatic hydrocarbon; PAH-EVs, EVs derived from PAH-treated hepatocytes; PYR, pyrene; ROS, reactive oxygen species; Th, thiourea; VitE, vitamin E; Wort, wortmannin.

Accepted

4 1. INTRODUCTION

Extracellular vesicles (EVs), i.e. membrane-enclosed vesicles released by cells into

the extracellular environment, are now recognized as important mediators of intercellular

communication [1]. Depending on the initial stimulus of their production, EVs are specifically

enriched in cargoes such as proteins, lipids, and nucleic acids that can be transferred into

recipient cells [2,3], thereby modulating cell functions. As liver is a multicellular organ

composed of hepatocytes and non-parenchymal cells, a role for EVs has emerged in the

pathogenesis of liver diseases, such as hepatitis, cirrhosis and hepatocellular carcinoma [4–6].

Thus, hepatocyte-derived EVs have been described to activate [7] or recruit [8,9] macrophages,

to activate stellate cells [10,11] and endothelial cells [12,13], thus leading to inflammation,

fibrogenesis and angiogenesis, processes well-known to participate in the development of liver

diseases.

Whereas oxidative stress is considered as a key event in the pathogenesis of liver

diseases [14–16], nothing is known about its possible induction in cells targeted by

hepatocyte-derived EVs. However, EVs isolated from some extrahepatic tissues have been reported to

generate reactive oxygen species (ROS) in recipient cells. Thus, the large EVs, i.e.

microvesicles, essentially isolated from circulating cells or from endothelial cells, have been

recently reviewed as ROS producers in recipient cells [17]. Besides, exosomes, the smaller type

of EVs, when released from skin keratinocytes [18], melanoma cell lines [19] or breast cancer

cells [20], were able to cause an increase in ROS levels in recipient cells. However, in all cases,

no proof of an oxidative damage was provided. Regarding cells producing EVs, a role for

oxidative stress in the EV release by pulmonary fibroblasts [21] or bronchial epithelial cells

[21,22] has already been shown, but nothing is known yet concerning hepatocytes.

Recently, for the first time in the field of toxicant-associated liver diseases, some

polycyclic aromatic hydrocarbons, widespread environmental contaminants, were

demonstrated to stimulate hepatocytes to release EVs with a modified composition and

Accepted

5

structure, as a mirror of mechanisms involved in PAH-induced cell death [23]. Interestingly,

PAHs have also been described to induce cell death by mechanisms that implicated oxidative

stress. For instance, benzo(a)pyrene (BP) can promote ROS production in liver cells both in

vitro [24,25] and in vivo [26], leading to oxidative damages such as lipid peroxidation in

cultured liver cells [27,28] or in mice liver [29], as well as DNA oxidation in hepatocytes [30]

or in rat liver [31]. Those oxidative damages were demonstrated to be responsible for cell death

in liver [27,28,32,33]. Pyrene (PYR) can also trigger ROS production in hepatocytes along with

a decrease in the expression of antioxidant enzymes, thus leading to a loss of viability [34]. In

mice, the depletion in the antioxidant glutathione was associated to PYR-induced liver damage

[35].

Although oxidative stress is a key component of the pathogenesis of liver diseases, no

role for hepatocyte-derived EVs has yet been described in the development of this process.

Therefore, we decided to investigate the involvement of EVs derived from hepatocytes treated

by PAHs in possible oxidative effects in healthy recipient hepatocytes. Particular attention was

also paid to the role of oxidative stress for their release. Using hepatocytes from both WIF-B9

cell line and primary rat cultures, we demonstrated that EVs derived from PAH-treated

hepatocytes were able to induce oxidative stress in recipient hepatocytes notably via an iron

enrichment of those EVs.

2. MATERIALS AND METHODS 2.1. Chemicals

Benzo(a)pyrene (BP), dibenzo(a,h)anthracene (DBA), pyrene (PYR),

N-acetyl-Asp-Glu-Val-Asp-7-amido-4 methylcoumarin (Ac-DEVD AMC), Hoechst 33342, dimethyl sulfoxide

(DMSO), cytochalasin B (CyB), methyl--cyclodextrin (MCD), wortmannin (Wort), bafilomycin A1 (BAF), desferrioxamine (DFO), thiourea (Th) and vitamin E (VitE) were all

purchased from Sigma–Aldrich (Saint Quentin Fallavier, France). SYTOX®Green was

Accepted

6

obtained from Life Technologies (Thermo Fisher Scientific Courtaboeuf, France) and

deferiprone (DFP) from Acros (Thermo Fisher Scientific Courtaboeuf, France).

Dihydroethidium (DHE), C11-BODIPY581/591, LysoTracker Red DND-99 and LysoSensor Yellow/Blue DND-160, MitoTracker red FM were obtained from Molecular Probes

(ThermoFisher Scientific, Courtaboeuf, France). Apocynin was acquired from Calbiochem

(Millipore, Saint-Quentin Les Yvelines, France). Mito-FerroGreen was obtained from Dojindo

(Munich, Europe).

Goat polyclonal anti-FTL (ferritin light chain) antibody was purchased from Origene

(Herford, Germany). Mouse monoclonal anti-HSC70, anti-gp91-phox and anti-p47-phox

antibodies were acquired from Santa Cruz Biotechnology (Heidelberg, Germany). Rabbit

monoclonal anti-ferritin heavy chain was purchased from Abcam (Paris, France). Horseradish

peroxidase (HRP)-conjugated secondary antibodies were from Dako (Agilent Technologies,

Courtaboeuf, France).

2.2. Cell isolation and culture

Two types of hepatocytes, primary rat hepatocytes and hybrid human/rat hepatocytes of

the WIF-B9 cell line, were studied. Primary hepatocytes were used as they are physiologically

and metabolically close to in vivo models. For longer incubation times, not suited for primary

hepatocyte survival, the WIF-B9 cell line was also tested due to its high level of differentiation

into hepatocytes, making it very sensitive to chemical toxicants, even at very low doses [36,37].

Primary rat hepatocytes. Adult hepatocytes were isolated from two-month-old male

Sgrague Dawley rats (Janvier labs, Le Genest-Saint-Isle, France) by perfusion of the liver with

a 17 µg/ml liberase (Roche Diagnostic, Meylan, France) solution to obtain liver dissociation.

All protocols (project APAFIS 6675-2016) were approved by our local ethic committee CREEA (Comité Rennais d’Ethique en matière d’Expérimentation Animale), and were in agreement with the European Union regulations concerning the use and protection of experimental animals (Directive

Accepted

7

2010/63/EU). Viability of the freshly isolated hepatocytes was routinely above 85%. Cells were seeded at densities of 1.5 x 106 cells in 35-mm petri dishes (Hoechst 33342 staining), 3 x 106 cells in 60-mm petri dishes (caspase-3/7 activity, EV isolation for NTA), 8 x 106 cells in 100-mm petri dishes (membrane fluidity determination and cholesterol measurement), and 15 x 106

cells in 150-mm petri dishes (EV isolation for protein, lipid, and EPR analysis). They were then cultured in a medium composed of 75% minimum Eagle’s medium and 25% medium 199 with

Hanks salts (Sigma-Aldrich, Saint-Quentin-Fallavier, France), supplemented with 10% fetal

calf serum (FCS) (Gibco, Illkirch, France) and containing 2.2 g/L NaHCO3 and 5 µg

streptomycin, 5 UI penicillin (Gibco, Illkirch, France), 3.2 µg bovine serum albumin (BSA)

(Eurobio, Courtaboeuf, France), and 5 µg bovine insulin (Sigma-Aldrich,

Saint-Quentin-Fallavier, France) per milliliter. Cells were incubated at 37°C under an atmosphere of 5% CO2

and 95% air. The medium was changed 3 h following seeding and replaced by the same medium

without FCS.

WIF-B9 cell line. WIF-B9 hybrid cells, obtained by fusion of Fao rat hepatoma cells

and WI-38 human fibroblasts, were a generous gift from Doris Cassio (UMR Inserm S757,

Université Paris-Sud, Orsay, France). WIF-B9 cells were grown in F12 Ham Coon’s nutrient

medium (Sigma-Aldrich, Saint-Quentin-Fallavier, France) supplemented with 5% FCS (Gibco,

Illkirch, France) and containing 2.2 g/L NaHCO3, 100 UI penicillin, 10 µg streptomycin,

and 0.25 μg amphotericin B (Sigma-Aldrich, Saint-Quentin-Fallavier, France) per

milliliter, 2 mM glutamine (Gibco, Illkirch, France), and HAT supplement [sodium

hypoxanthine (10 µM), aminopterin (40 nM), and thymidine (1.6 µM)] (Gibco, Illkirch,

France). Cells were incubated at 37°C in an atmosphere of 5% CO2 and 95% air. Cells were

seeded at 12.5 x 103 cells/cm2 and cultured for seven days, until reaching approximately 80% of confluence, before PAH treatment. Before treatment, the medium was changed and replaced

Accepted

8

by the same medium but previously deprived of EVs. EV depletion was obtained by

ultracentrifugation of the medium containing 25% FCS, at 100,000 x g for 14 h at 4 °C.

2.3. Cell treatment

For this study, we selected the same PAHs as for our previous work [23] i.e. BP, DBA,

and PYR based on their variable presence in food [38] and affinity for the aryl hydrocarbon

receptor (AhR) [39]. In order to isolate EVs from PAH-treated hepatocytes, cells were exposed

to 100 nM of each PAH, or 0.0005 % dimethyl sulfoxide (DMSO) as control cultures, for 18 h

for primary rat hepatocytes and 72 h for WIF-B9 hepatocytes. This PAH concentration was

chosen based on results previously published [23]. Then, to test the impact of those EVs on

healthy recipient hepatocytes, a preliminary study of dose and time response on cell death was

conducted in both cell types. For this purpose, they were exposed to EVs at various

concentrations (2.5, 5 or 15 µg of EV proteins/mL) for various incubation times (5, 24, 48 or

72 hours for WIF-B9, and 5 or 18 hours for primary rat hepatocytes). Those EV concentrations

roughly corresponded to that found in culture media of WIF-B9 hepatocytes treated for 72 hours

by PAHs or of primary rat hepatocytes exposed during 18 hours. As the levels of apoptosis

induced by EVs derived from PAH-treated WIF-B9 and primary rat hepatocytes were the

highest at 24 hours and 18 hours, respectively (Supplementary Figure S1), the following

experiments were performed using those exposure times. Concerning EV concentration, a

highest apoptosis was obtained with the 5 and 15 µg/mL concentrations leading us to keep the

5 µg/mL concentration for the following experiments. In the experiments aimed at investigating

the mechanisms underlying the effect of EVs on recipient hepatocytes, cultures were pre-treated

with various compounds (Supplementary Table S1) for 30 min prior to co-exposure with the

toxicants.

Accepted

9 2.4. EV analysis

EV isolation from PAH-treated hepatocytes. After each treatment with toxicant,

culture media were collected to isolate total EVs. The media were first centrifuged at 3,650 x g

for 10 min to remove any cells, cell debris and apoptotic bodies. Total EVs were then pelleted

by ultracentrifugation (Optima L-90 K Ultracentrifuge, Beckman Coulter, Villepinte, France)

of the cleared supernatants at 100,000 x g for 1 h 45 min at 4°C in a SW28-1 swinging-bucket

rotor (Beckman Coulter, Villepinte, France). Pellets were then washed with PBS and

re-centrifuged at the same speed. Finally, EVs were re-suspended in 100 to 200 µL PBS,

depending on the quantity.

Protein concentration. The protein concentrations of cell lysates and EVs were

determined by the Pierce BCA protein assay kit (Thermo Fisher Scientific, Illkirch, France)

using BSA as the standard. Hepatocytes were lysed in RIPA buffer (50 mM Tris buffer, 150

mM NaCl, 0.1 % SDS, 1 % Triton X-100, 0.5 % sodium deoxycholate, 50 mM NaF, 5 mM EDTA pH 8) supplemented with protease inhibitors (1 mM orthovanadate, 1 mM

phenylmethylsulfonyl fluoride, 5 µg/ml leupeptin, 0.1 µg/ml aprotinin and 0.5 mM

dithiothreitol) before sonication on ice.

Protein expression. Equal amounts of protein (at least 10 µg) were denatured for 5 min

at 95 °C in RIPA buffer and then separated by sodium dodecyl sulfate (SDS) polyacrylamide

gel electrophoresis (SDS–PAGE) using 12% or 15% polyacrylamide gels. After overnight

transfer onto nitrocellulose membranes (Millipore, St Quentin-en-Yvelines, France) and

blocking for 1 hour at room temperature with a TBS solution containing 5% BSA, primary

antibodies were added and the membranes incubated for 2 hours at room temperature

[anti-gp91-phox (1:100), anti-p47-phox (1:1000), anti-FTL (1:1000), and anti-HSC70 (1:10000)].

Then, membranes were washed three times with 0.1% Tween 20 in TBS for 10 min and

incubated in blocking buffer containing appropriate horseradish peroxidase-conjugated

secondary antibodies (1:2500) for 1 hour at room temperature. Following incubation, they were

Accepted

10

washed six times with 0.1% Tween 20 in TBS for 5 min. Immunolabeled proteins were then

visualized by chemiluminescence using an enhanced chemiluminescence (ECL) solution

(Tris-HCl 0.1 M pH 8.5, 0.22 mM coumaric acid, 1.25 mM luminol, and 0.009 % H2O2) and a

ChemiDoc™ XRS+analyzer, and the images processed using Image Lab 6.0 (Bio-Rad,

Marnes-la-Coquette, France). HSC70 primary antibodies were used to evaluate protein loading.

Independent experiments could not be run on the same gel. Thus, the relative density

ratio was calculated. First, band densities of the protein of interest (FTL, gp91-phox, p47-phox)

and that of the loading controls (Hsc70) were measured on every blot and the backgrounds

subtracted. Then, the density of the protein of interest in each lane was divided by the respective

density of the loading control on every blot. Finally, each normalized density of the protein of

interest for the EVs was divided by the normalized density of the control EVs (first lane on the

blots). Each value for the control EV from independent blots was set to one, as the density ratios

were systematically relative to the control EV ratio. Differences in the means of relative density

ratios allowed us to compare expression levels of the proteins of interest from different blots.

For multiple bands, the densities of each band were added and the sum used as the density of

the protein of interest.

Cryo-electron microscopy. Vitrification of EVs was performed using an automatic

plunge freezer (EM GP, Leica) under controlled humidity and temperature [40]. The samples

(i.e. 10 µL of EVs at a concentration of 1012 _{particles/mL) were deposited on glow-discharged}

electron microscope grids followed by blotting and vitrification by rapid freezing into liquid

ethane. Grids were transferred to a single-axis cryo-holder (model 626, Gatan) and were

observed using a 200 kV electron microscope (Tecnai G2 T20 Sphera, FEI) equipped with a 4k

× 4k CCD camera (model USC 4000, Gatan). Micrographs were acquired under low electron

doses using the camera in binning mode 1 and at a nominal magnification of 25,000 X.

Accepted

11

Iron measurement. Five microliters of the EV suspension were incubated for 20 min

on formvar carbon-coated grids (Nickel Grid 300 mesh, Agar Scientific, Oxford Instruments,

Gometz-la-Ville, France). Excess liquid was removed by blotting and the grid was then

negatively stained with 2% uranyl acetate followed by blotting to remove excess liquid.

Samples were imaged using a Jeol JEM 2100 HR microscope (JEOL Ltd, JEOL Europe SAS,

Croissy-sur-Seine, France) operated at an acceleration voltage of 200 kV, equipped with a

Gatan Orius SC200D camera (Gatan Inc) and coupled with an Energy Dispersive X-Ray

Spectroscopy (EDS) analyser EDX Oxford X-Max 80T that characterizes the elemental

composition.

EV uptake by recipient cells. EVs were first labelled using the PKH67 Green

Fluorescent Cell Linker Kit (Sigma–Aldrich, Saint Quentin Fallavier, France). Briefly, following the manufacturer’s instructions, the washed EV pellets were resuspended in 250 µL

of Diluent C and then mixed with 250 µL of a solution of 4 µM PKH67 during 4 min. Then,

300 µL 1 % BSA were added to bind the excess PKH67 dye. Finally, the EVs were

ultracentrifuged at 100 000 x g during 1 h 45 min, and washed twice by ultracentrifugation.

Samples without EVs or, with EVs but without PKH67, were also prepared following the same

steps to be used as negative controls. For microscopy observation of EV uptake by hepatocytes,

cells were seeded on glass coverslips and treated with 5 µg/mL PHK67 labelled-EVs

(PKH67-EVs) during 24 hours. After treatment, cells were fixed in 4 % paraformaldehyde for 20 min,

washed with PBS 1X. Slides were then viewed using a confocal fluorescence microscope

LEICA DMI 6000 CS (Leica Microsystems, Wetzlar, Germany) equipped with a 63× objective.

A quantification by spectrophometry was also performed. Approximatively 1.5 million of

hepatocytes were treated with 5 µg/mL PHK67-EVs for 24 hours. After treatment, cells were

washed in PBS, centrifuged 5 min at 800 x g, resuspended in 200 µL PBS and finally lyzed by

sonication. Two hundred microliters of cell lysates were then transferred to 96-well plate and

Accepted

12

the PKH67 fluorescence was measured at 502 nm by spectrophotometry using a Perkin Elmer

Enspire 2300 multilabel reader (Perkin Elmer, USA). PKH67 fluorescence in cells was

normalized by the protein quantity.

2.5. Cell death evaluation of recipient hepatocytes

After treatment by hepatocyte-derived EVs, cells were tested for both apoptotic and

necrotic cell death. Apoptotic cells were identified by chromatin condensation and

morphological changes in the nucleus using the blue-fluorescence chromatin dye Hoechst

33342, while necrotic cells were evidenced by the SYTOX®Green nucleic acid stain, only

permeant to cells with compromised plasma membranes. Briefly, WIF-B9 and primary rat

hepatocytes were stained at 37 °C with 175 nM SYTOX®Green and also with 50 µg/mL and

100 µg/mL Hoechst 33342 for 20 and 30 min, respectively. Apoptotic and necrotic cells were

then counted using a fluorescence microscope (Axio Scope A1, ZEISS, Marly le Roi, France);

400 cells in total were examined for each condition.

In order to further identify cell death mode, the caspase-3/7 activity assay was

performed. Cells were lysed in the caspase activity buffer as previously described [27]. Eighty

micrograms of protein were thus incubated with 80 μM DEVD-AMC for 2 hours at 37 °C.

Caspase-mediated cleavage of DEVD-AMC was evaluated by spectrofluorimetry (Spectramax

Gemini; Molecular Devices, San Jose, California, United States) using 380 nm excitation and

440 nm emission wavelengths.

2.6. Determination of organelle alteration

Electron microscopy of recipient hepatocytes. Ultrastructure changes of recipient

WIF-B9 cells were visualized by transmission electron microscopy. After 24 hour-exposure to

EVs, hepatocytes were rinsed with 0.15M cacodylate buffer, pH 7.4 and fixed by drop-wise

Accepted

13

addition of glutaraldehyde (2.5%) in 0.15M cacodylate buffer, for 1 hour. They were then

washed with 0.15M cacodylate buffer and postfixed with 1.5% osmium tetroxide solution

containing 1.5% potassium ferrocyanide for 1 hour. Samples were first washed with cacodylate

buffer then with water, stained with 1% uranyl acetate for 1 hour and washed again. Next,

samples were dehydrated with graded alcohol series following standard procedure prior to

inclusion in Epon-Araldite-DMP30 resin (polymerized at 60°C for 48 hours). Sections (0.5 µm)

were cut on a Leica UC7 microtome (Leica Microsystems, Wetzlar, Germany) and stained with

toluidine blue. Ultrathin sections (80 nm) were obtained and mounted onto copper grids.

Sample examination was performed with a JEOL 1400 transmission electron microscope (Jeol

Co, Tokyo, Japan) operated at 120 kV and images were digitally captured with an Orius 1000

Gatan Camera (Gatan Ametek, Pleasanton, USA).

Fluorescence microscopy of lysosomes. Lysosomes were labeled by LysoTracker red,

a red acidotropic fluoroprobe. After treatment with EVs, hepatocytes, seeded on glass

coverslips, were incubated with 1 µM LysoTracker in the culture medium, at 37 °C for 1 h

(WIF-B9 hepatocytes) or for 10 min (primary rat hepatocytes). After labelling, cells were fixed

in 4 % paraformaldehyde for 20 min, washed with PBS. Slides were then viewed using a

confocal fluorescence microscope LEICA DMI 6000 CS (Leica Microsystems, Wetzlar,

Germany) equipped with a 40X objective.

Measurement of lysosomal pH. Adapted from a previous paper [41], changes in

lysosomal pH were analyzed using the fluorescent ratiometric probe LysoSensor Yellow/Blue

DND-160, which exclusively localizes within the lysosomes. The LysoSensor dye emits blue

fluorescence in neutral environments, but switches to yellow fluorescence in more acidic

environments. After treatment with EVs, hepatocytes, seeded on a 96-well plate, were incubated

with 1 µM LysoSensor for 2 min at 37°C. As a positive control of lysosomal alkalinization,

some hepatocytes were treated, for 30 min prior to LysoSensor incubation, with 50 nM

Accepted

14

bafilomycin A1, a known inhibitor of lysosomal H+-ATPase (also known as V-ATPase).In parallel, a pH calibration curve was generated. Cells were incubated 2 min at 37°C with 1 µM

LysoSensor and then 10 min at 37°C with buffered solutions at pH 4, 5, 6 or 6.5 containing 125 mM KCl, 25 mM NaCl, 10 μM monensin, 10 µM nigericin and 25 mM 2-[N-morpholino]

ethanesulfonic acid (MES). Finally, using a Perkin Elmer Enspire 2300 multilabel reader (Perkin Elmer, USA), fluorescence was measured at 440 nm and 540 nm.

2.7. Oxidative stress measurement in hepatocytes

ROS production. Intracellular ROS production was evaluated using dihydroethidium

(DHE), as previously described [33]. Briefly, after treatment by hepatocyte-derived EVs,

recipient cells were incubated with 25 μM DHE for 1 h at 37°C. Fluorescence (Ex 396 nm/Em

580 nm) was recorded using a microplate reader (EnSpire Multimode 2300 Plate Reader, Perkin

Elmer, Waltham, United States). Results were given as relative fluorescence units (RFU)/g protein.

Analysis of lipid peroxidation. Lipid peroxidation was measured using the fluoroprobe

C11-BODIPY581/591 _{(4,4-difluoro-5- [4-phenyl-1,3-butadienyl] -4-}

bora-3a,4a-diaza-s-indacine-3-undecanoic acid), as previously described [28]. C11-BODIPY581/591, which

incorporates into cell membranes, shifts from red to green fluorescence upon oxidation. Briefly,

cells were incubated with 10 µM C11-BODIPY581/591 for 1 h at 37°C. Then, fluorescence was recorded by a SpectraMax Gemini spectrofluorimeter (Molecular Devices Sunnyvale, United

States) using two pairs of wavelengths to measure the amount of reduced (ex 590 nm, em 635 nm) and oxidized probe (ex 485 nm, em 535 nm). Lipid peroxidation was then quantified by calculating the ratio of oxidized probe/ total probe (oxidized + reduced probe).

Low-molecular-weight iron content. Measurement of intracellular low molecular

weight (LMW) iron was based upon the capacity of deferiprone to chelate only LMW iron and

Accepted

15

to give a paramagnetic chelate, which can be directly detectable by electron paramagnetic

resonance in whole hepatocytes, as previously described [42].

Mitochondrial free iron. Free iron content specifically located in mitochondria was

detected using the cell-permeable fluorescent probe Mito-FerroGreen, while mitochondria

themselves were visualized by staining with the fluorescent probe MitoTracker red FM.

After treatment with EVs, hepatocytes seeded on glass coverslips were washed with

PBS and then incubated with 50 nM MitoTracker red FM and 5 µM Mito-FerroGreen in PBS,

at 37 °C for 30 min. After labelling, cells were washed with PBS. Slides were then viewed using

a confocal fluorescence microscope LEICA DMI 6000 CS (Leica Microsystems, Wetzlar,

Germany) equipped with a 40X objective. Mitochondrial iron content was quantified, cell by

cell, by the relative ratio of green over red fluorescence; 100 cells in total were examined for

each condition.

2.8. Cytochrome P450 1 activity.

CYP1 activity was estimated by the ethoxyresorufin O-deethylase (EROD) assay that

relies on the conversion of ethoxyresorufin into resorufin by CYP1 enzymes. Rapidly, after

treatment, cells were incubated with 1.5 mM salicylamide (inhibitor of phase II-conjugating enzymes) and 5 μM ethoxyresorufin. Fluorescence of resorufin (excitation at 544 nm and

emission at 584 nm) was monitored for 30 min at 37°C using a microplate reader (EnSpire

Multimode 2300 Plate Reader, Perkin Elmer, Waltham, United States). EROD activity was

expressed as pg resorufin per min and mg protein.

2.9. Statistical analysis

Values are presented as means ± standard deviation from at least three independent

experiments. Statistical analyses were performed using one-way analysis of variance (ANOVA)

Accepted

16

followed by a Newman-Keuls post-test. Significance was accepted at p<0.05. All statistical

analyses were performed using GraphPad Prism5 software (San Diego, United States).

3. RESULTS

3.1. PAH treatment of hepatocytes induces a release of EVs depending on oxidative stress.

BP was previously reported as capable of increasing ROS generation and lipid

peroxidation in liver epithelial cells [28], primary rat hepatocytes [27] and WIF-B9 hepatocytes

[33] leading to cell death. Therefore, we decided to evaluate the ability of the three selected

PAHs to generate oxidative stress in WIF-B9 and primary rat hepatocytes. As expected, BP

treatment caused an increase of ROS production and consequent lipid peroxidation after 72

hours of exposure in WIF-B9 hepatocytes (Figure 1A) or after 18 hours in primary rat

hepatocytes (Figure 1B). Similar results were obtained with DBA and PYR (Figures 1A and

1B). Using thiourea, a ROS scavenger, and vitamin E, a free radical chain breaking antioxidant,

oxidative stress was demonstrated to be involved in PAH-induced cell death (Figures 1C and

1D).

Given that oxidative stress has been demonstrated as inducing EV release in order to

protect the producer cells from oxidative damages [43], we examined the effect of antioxidants,

i.e. thiourea or vitamin E, on the EV release by PAH-treated hepatocytes. As described in

Materials and Methods, EVs were isolated from the cell culture media by differential

ultracentrifugation and quantified by NTA. Antioxidants significantly reduced the increase in

EV release induced by all PAHs (Figures 1E and 1F), thus indicating a role for PAH-induced

oxidative stress in EV release by hepatocytes.

Accepted

17

3.2. EVs derived from PAH-treated hepatocytes are enriched in proteins related to oxidative stress.

As oxidative stress was involved in the release of EVs upon PAH exposure, this release

was proposed as a cellular protective mechanism against oxidative stress by taking away

deleterious molecules. Therefore, we decided to look for several proteins related to oxidative

stress in EVs [43–46]. First, we examined the changes in the expression levels of NADPH

oxidase subunits, gp91-phox (Figures 2A and 2B) and p47-phox (Figures 2C and 2D). Both

NADPH oxidase subunits were higher in EVs derived from PAH-treated hepatocytes

(PAH-EVs) than in EVs derived from untreated hepatocytes (CTRL-(PAH-EVs). It is noteworthy that two

bands were obtained for the p47-phox regulatory subunit. This could be ascribed to the

phosphorylation of this subunit, thus suggesting a stronger activation of NADPH oxidase in

PAH-EVs. Second, we explored protein expression of a subunit of ferritin, a storage protein of

low-molecular-weight iron (LMW iron). This iron species is known to trigger oxidative stress

by catalyzing the formation of a highly reactive free radical, the hydroxyl radical, via Fenton

or Haber-Weiss reaction. Expression of both ferritin subunits, the ferritin light chain (FTL) and

the ferritin heavy chain (FTH), was found to be markedly increased in PAH-EVs from both

WIF-B9 and primary hepatocytes (by approximately 60-70 % for FTL, 30-60 % for FTH from

WIF-B9 hepatocytes and 100 % for FTH from primary hepatocytes) (Figures 2E-2H). Then,

observing by cryo-electron microscopy the ultrastructure of EVs derived from WIF-B9

hepatocytes, large dense structures were found in EVs (Figure 3A, white arrows). These

structures may be assimilated to large protein complexes and have been suggested by

Truman-Rosentsvit (2018) as to be typical ferritin-iron cores [47]. Interestingly, PAH-EVs contained

more of those structures (Figure 3A), thus confirming the results obtained by western-blotting.

In order to test a possible elevation of iron within the PAH-EVs, we measured iron atom levels

in EVs by electron microscopy coupled with an Energy Dispersive X-Ray Spectroscopy (EDS)

Accepted

18

analyser. This technique allows a characterization of the elemental composition of an analysed

volume. Levels of iron atom were increased in PAH-EVs compared to CTRL-EVs (Figure 3B).

It is noteworthy that iron atom was the only one with significantly modified levels in EVs when

hepatocytes were treated by PAHs (Supplementary Figure S2). Interestingly, alongside the

release of ferritin in EVs, a decrease in LMW iron was detected at 96 hours of exposure to BP,

after a transient increase at 24 hours (Supplementary Figure S3). Those results collectively

illustrated that PAH-EVs were enriched in iron most probably stored in ferritin complexes.

3.3. EV released by PAH-treated parent hepatocytes can induce oxidative stress and related damages in recipient hepatocytes

Due to their content in pro-oxidant components such as NADPH oxidase and iron,

PAH-EVs may transfer them in recipient hepatocytes. NADPH oxidase generates superoxide anion

radical that can lead to the formation of hydrogen peroxide. Iron catalyzes the Fenton reaction

that used hydrogen peroxide to produce the highly oxidant hydroxyl radical. Therefore, we

wondered whether PAH-EVs could promote oxidative stress and hence harmful processes in

recipient hepatocytes. For this purpose, WIF-B9 or primary rat hepatocytes were treated by

PAH-EVs in comparison with CTRL-EVs for 24 or 18 hours, respectively. CTRL-EVs did not

induce any significant changes in ROS levels compared to hepatocytes without treatment by

EVs (Figure 4A and 4B). At the opposite, PAH-EVs increased the ROS levels compared to

CTRL-EVs both in WIF-B9 and primary rat hepatocytes (Figures 4A and 4B), thus resulting in

a lipid peroxidation of recipient hepatocytes (Figures 4C and 4D).

Consequently, we decided to search for oxidative damages of hepatocytes such as

mitochondrial alteration and cell death. In WIF-B9 hepatocytes, mitochondrion ultrastructure

was observed by transmission electron microscopy. When compared to CTRL-EV exposure,

PAH-EV treatment induced a swelling of mitochondria (Figure 5) and an overall increase in

Accepted

19

mitochondrion size (Supplementary Table S2). Mitochondria would also undergo cristae

alterations (Figure 5, white arrows) as well as fusion (Figure 5, white asterisks) and fission

(Figure 5, solid black arrows) processes. These observations indicated alterations of

mitochondria structure, suggestive of a possible ongoing cell death process.

Cell death was therefore estimated by Hoechst 33342 and Sytox®Green staining in

order to determine by fluorescent microscopy the number of apoptotic and necrotic hepatocytes.

Compared to untreated hepatocytes, cells exposed to CTRL-EVs did not exhibit any significant

increase in apoptosis. In contrast, PAH-EVs led to a significant increase (by  100%) in the number of apoptotic cells for both WIF-B9 (Figure 6A) and primary rat (Figure 6B)

hepatocytes. Apoptotic cell death was accompanied by a significant increase in caspase 3/7

activity in both hepatocyte types (by  50%) (Figures 6C and 6D). Neither CTRL-EVs nor PAH-EVs changed the number of necrotic cells (data not shown). Finally, using the

antioxidants, i.e. thiourea and vitamin E, PAH-EVs-induced cell death was prevented (Figures

6E and 6F), suggesting a role for oxidative stress in this process.

3.4. PAH-EV-induced apoptosis of recipient hepatocytes is dependent on EV endocytosis and consecutive lysosome permeabilization

In order to understand how PAH-EVs could induce oxidative stress and hence apoptosis

of recipient hepatocytes, we first decided to determine whether EVs were internalized by

recipient cells. Indeed, uptake of EVs is generally necessary to obtain an effect on recipient

cells [48]. The internalization of CTRL-EVs was tested after labelling of EVs with the lipophilic

fluorochrome PKH-67. A transfer of PKH67 green fluorescence was observed indicating the

EV incorporation within the recipient hepatocytes (Figure 7A, left panel). We next quantified

spectrophotometrically this PKH67-EV uptake. Thus, it appears that uptake of PAH-EVs was

significantly increased as compared to CTRL-EVs (+50 %) (Figure 7A, right panel). Note that

Accepted

20

the fluorescence of PKH67-EV was more specifically localized around the nucleus, thus

suggesting an EV presence in lysosomes and hence a possible EV uptake via the intracellular

vesicular traffic along the endocytic route (Figure 7A, left panel).

Therefore, in order to check the possibility that EV could reach lysosomes, we

investigated various mechanisms of EV engulfment by WIF-B9 hepatocytes; to do so, we used

cytochalasin B (CyB) and wortmannin (Wort), two inhibitors of endocytosis and phagocytosis,

or methyl-β-cyclodextrin (MCD), an inhibitor of a particular endocytosis dependent on lipid

rafts. This latter endocytosis type was investigated because lipid rafts (i.e. cholesterol-enriched

microdomains) have been shown to be involved in BP-induced apoptosis of liver epithelial cells

[49] and of primary rat hepatocytes [27]; also an increase in cholesterol content has been

previously described in PAH-EVs [23].

In this study, MCD affected neither CTRL-EV nor PAH-EV uptake (Figure 7A, right

panel), thus ruling out any possible implication of lipid rafts in EV uptake. In contrast, the

reduction in uptake of both CTRL-EVs and PAH-EVs by CyB and Wort (Figure 7A, right

panel) highlighted the involvement of endocytosis or phagocytosis in this process. Finally, the

same compounds were also tested on PAH-EV-induced apoptosis of recipient hepatocytes.

Similarly, only CyB and Wort protected from apoptosis both WIF-B9 (Figure 7B) and primary

rat (Figure 7C) hepatocytes, thus suggesting the key role of endocytosis or phagocytosis. These

mechanisms of EV uptake therefore supported our assumption that hepatocyte-derived EVs

could reach lysosomes within recipient hepatocytes following the classical vesicular traffic via

endosomes.

We then hypothesized that the accumulation of EVs in lysosomes could alter those

organelles leading to cell death. Using bafilomycin A1 (BAF), a selective inhibitor of

V-ATPase, in order to reduce lysosome acidity and hence lysosomal hydrolase activity, the

PAH-EVs-induced apoptosis of recipient hepatocytes was significantly decreased both in WIF-B9

Accepted

21

(Figure 7B) and primary rat (Figure 7C) hepatocytes. In addition, bafilomycin was able to

protect from mitochondrial damages (Supplementary Figure S4). Thus, lysosomes appeared as

a key organelle in the development of PAH-EV-induced cell death. In this context, we

questioned about possible alterations of lysosomes when hepatocytes were exposed to

PAH-EVs. First, the red fluorescent probe Lysotracker Red DND-99, which specifically localizes in

lysosomes, was used in both WIF-B9 (Figure 8A) and primary rat (Figure 8B) hepatocytes.

PAH-EVs caused an increase in lysosome size or a diffusion of fluorescence throughout the

whole cell, reflecting a possible lysosome membrane permeabilization in recipient hepatocytes

(Figures 8A and 8B). CTRL-EVs do not alter lysosome fluorescence unlike PAH-EVs. Second,

to support those observations, lysosome ultrastructure of WIF-B9 hepatocytes was examined

by transmission electron microscopy. After PAH-EVs treatment, an increase in lysosome size

was also observed (Figure 8C) and measured (Supplementary Table S3). Additionally,

hepatocytes exposed to PAH-EVs displayed lysosome membrane permeabilization

characterized by a rupture of the lysosome membrane and a visible leak of content (Figure 8C,

white arrows).

Since alteration of lysosome membranes have been reported to affect V-ATPase and

hence intraluminal pH [50], we measured lysosome pH using the LysoSensor Yellow/Blue

DND-160 probe. A 0.2 increase in pH was found (Supplementary Figure S5) and ascertained

the alteration of lysosome membrane.

Overall, these results demonstrated that PAH-EVs targeted lysosomes causing a

membrane permeabilization and subsequent cell death.

Accepted

22

3.5. The pro-oxidant components of EVs, NADPH oxidase and iron, could be involved in PAH-EV-induced apoptosis of recipient hepatocytes

Since PAH-EVs are enriched in proteins related to oxidative stress, such as NADPH

oxidase and ferritin, their accumulation in lysosomes of recipient hepatocytes could participate

to the alteration of those organelles and hence to apoptosis. To test this hypothesis, we first used

two LMW-iron chelators, i.e. desferrioxamine and deferiprone. Whereas deferiprone chelates

iron in all cell compartments, desferrioxamine is endocytosed and transported to lysosomes,

making it a lysosome-specific iron chelator. The PAH-EVs-induced death of recipient WIF-B9

and primary rat hepatocytes was significantly attenuated by deferiprone or desferrioxamine

(Figures 7B and 7C). Additionally, the use of either chelator on recipient WIF-B9 hepatocytes

limited the PAH-EV-induced lysosome membrane permeabilization (Supplementary Figure

S6). Finally, BP-EVs also significantly increased free iron in mitochondria (Figure 9); such an

increase was reduced by the lysosome-specific iron chelator, desferrioxamine. Logically,

desferrioxamine also protected from PAH-induced mitochondria alterations (Supplementary

Figure S4). Altogether, these results suggested that LMW-iron overload in lysosomes is

implicated in alterations of those organelles consequently leading to mitochondrial damages

and cell death. In line with this, inhibition of NADPH oxidase by apocynin prevented the

apoptosis induced by exposure to PAH-EVs in recipient hepatocytes (Figures 7B and 7C).

As BP has already been described to induce oxidative stress and lysosome

permeabilization by its own metabolism [51] [27] , one might argue that the above effects could

be due to the presence of PAHs in EVs, thereby triggering oxidative stress in the recipient

hepatocytes. In order to discard such a mechanism, we measured EROD activity, as an indicator

of CYP1 activity. Indeed, the CYP1 enzymes, that are involved in the metabolism of BP and

DBA, are known to exhibit a high increase in activity in BP- or DBA-treated hepatocytes due

to an elevation of their mRNA expression [23]. Here, CYP1 activity in recipient hepatocytes

Accepted

23

was not increased upon exposure to PAHs-EVs (Supplementary Figure S7), thus suggesting

that PAHs were not, or in a very negligible quantity, transferred to recipient cells through EVs.

4. DISCUSSION

A growing body of evidences supports the major role of EVs as players of cell

communication for the pathogenesis of liver diseases. Yet, nothing is known about their

possible involvement in the onset of a key process in the course of liver diseases, namely

oxidative stress of the parenchymal hepatic cells. Here, we report, for the first time in

hepatocytes and also in the context of exposure to environmental contaminants, that EVs are

able to trigger oxidative damages in healthy recipient cells through a requisite internalization

of EVs. Thus, the only two published studies, about exposure of hepatocytes to EVs and a

possible related oxidative stress, have limited their results to the evaluation of increased ROS

levels [52,53]. In our work, in addition to ROS production in recipient hepatocytes, exposure

to EVs was demonstrated to trigger lipid peroxidation that in turn caused cell death. In addition,

as previously reported for ethanol [54] or ischemia-reperfusion [53], PAHs, by inducing

oxidative stress in the producer hepatocytes, promoted EV release. This therefore suggests that

oxidative stress is critical, in the context of PAH hepatotoxicity, for both EV release and

EV-induced cell death.

In order to understand the mechanisms by which only PAH-EVs could induce oxidative

damages in recipient hepatocytes, the specific content of these EVs with regards to compounds

related to oxidative stress was compared to CTRL-EV content. Indeed, PAH-induced oxidative

stress was clearly identified as the trigger of EV release by the parent hepatocytes. Such an

oxidation within the producer cell has been described to promote EV release by various

mechanisms [43,55]. When studied, such a release is recurrently proposed as a protective

mechanism to discard harmful molecules such as oxidized proteins, mtDNA, phospholipids and

Accepted

24

ROS production systems [43–46]. In our present work, NADPH oxidase, an enzyme that

produces the free radical superoxide anion, was shown, for the first time in hepatocyte-derived

EVs, to be more strongly expressed when producer hepatocytes exhibited an oxidative stress.

Regarding other physiopathological conditions such as sepsis or inflammation, an increase in

NADPH oxidase expression has also been reported in EVs isolated from bloodstream cells,

which also exhibited an elevation of ROS levels [56,57]. Taken altogether, those results thus

further support the possible role of EVs to protect the producer cells from oxidative stress. It is

noteworthy that Gambim (2007) and Janizewski (2004) also showed that the NADPH oxidase

of EVs was functional [56,57].

Expression of another protein related to oxidative stress, namely ferritin, was

investigated in PAH-EVs. Recently, it was demonstrated that an important pathway for

iron-loaded ferritin secretion occurred through the specialized smaller EV types, that is, exosomes

[47]. In our models, exosomes were previously shown to be the main EV type [23]. The ferritin

protein complex composed of 24 heavy and light subunits is able to store up to 4500 atoms of

the iron element in a nontoxic form from which iron could be released to trigger oxidative

damages. Here, the ferritin light subunit was found to be more expressed in PAH-EVs than in

CTRL-EVs. In contrast to NADPH oxidase expression in EVs, such an elevation of ferritin

light subunit has already been described in EVs released by stressed primary rat hepatocytes

[58]. However, so far, authors have not checked a possible co-precipitation of free ferritin with

EVs during their ultracentrifugation isolation procedure. In our study, cryo-electron microscopy

ascertained the increased presence in PAH-EVs of large dense structures, similar to the typical

ferritin-iron cores identified in macrophage EVs [47]. Furthermore, for the first time in EVs

released by cells exposed to oxidative stress, an elevation of the levels of iron atom was

uncovered by EDS microscopy in PAH-EVs, in accordance with the increase in expression of

the iron storage protein ferritin. This supports the assumption by Truman-Rosentsvit (2018) that

Accepted

25

secretion of ferritin via exosomes could be an export of ferric iron [47]. This release of iron via

EVs could be related once again to a mean for the hepatocytes to discard harmful molecules

capable of inducing oxidative stress and cell death. Indeed, Brown et al. (2019) [59] have

evidence a possible resistance towards ferroptosis, a new cell death type recently considered

for NASH [60], by the way of an iron expulsion in EVs, more specifically in exosomes.

Ferroptosis is an iron-dependent, necrosis-like programmed cell death characterized by the

accumulation of lipid reactive oxygen species, and hence inhibited by lipophilic antioxidant

[61,62]. In our model of WIB-B9 hepatocytes exposed to BP, a transient increase in LMW iron

levels has been demonstrated, by electron resonance spectroscopy, at an early time point (24 h)

followed by a significant decrease later on (96 h). In addition, the lipophilic antioxidant, vitamin

E inhibited the release of EVs. Both those results lead us to suggest that an onset of ferroptosis

might be triggered by BP, but would be stopped by the release of exosomes containing

iron-loaded ferritin. To sum up, it appears that hepatocytes when submitted to oxidative stress, can

release harmful species via EVs in order to protect themselves.

However, EVs loaded with such harmful species could then be engulfed by neighboring

hepatocytes. Regarding that point, hepatocytes exposed to PAH-EVs were thus demonstrated

to sustain oxidative stress leading to cell death, provided that those EVs were internalized in

the recipient hepatocytes. This uptake was more important with PAH-EVs than CTRL-EVs.

This could be ascribed to their difference in composition formerly reported, notably the increase

in cholesterol of PAH-EVs [23]. Indeed, a cholesterol enrichment of surfaces can promote cell

adhesion [63,64]. Binding of EVs to recipient cells is now recognized to facilitate subsequent

endocytosis [48,65]. EVs derived from hepatocytes submitted to other types of stress such as

heat shock [66] or ethanol exposure [54] have also been described to be internalized, but without

indicating the prerequisite of an EV engulfment to obtain inflammation or apoptosis of the

recipient hepatocytes, respectively. This requirement of an EV uptake is actually decisive to

Accepted

26

explain why PAH-EVs but not CTRL-EVs were able to induce oxidative stress in recipient

hepatocytes. Indeed, when endocytosed, EVs containing harmful cargoes, can go along the

classical pathway of intracellular vesicular traffic to reach lysosomes [65,67]. Up to now, the

only relationship between lysosomes and EVs, reported in literature, has dealt with a possible

increase in EV release due to lysosome alteration [68,69]. Here, the accumulation of PAH-EVs

in lysosomes led to a structural alteration of lysosomes, especially a lysosomal membrane

permeabilization, an event well-known as being involved in several types of cell death,

including apoptosis [50]. To our knowledge, it is the first time, thanks to the use of bafilomycin

A1, an inhibitor of V-ATPase and hence of hydrolase activity, that EVs were found to trigger

cell death via lysosomal alteration. It is noteworthy to consider that protease leakage from

lysosome [50] and lysosomal alkalinization [70], can induce downstream mitochondrial

damages, but also ROS generation and caspase 3/7 activation. Thus, we also demonstrate for

the first time that EVs, notably PAH-EVs, can induce major structural damages of mitochondria

such as swelling and cristae alterations. These results are in accordance with the increase in

mitochondrial permeability previously found in hepatocytes treated by circulating EVs from

injured mice [53]. Interestingly, mitochondrial fusion and fission, considered as adaptive

responses towards severe mitochondrial stress [71,72], might possibly occur in hepatocytes

treated by PAH-EVs. Those processes, notably fission, are known to lead to apoptosis

depending on the context [73]. It is also interesting to emphasize that, in a context of excessive

production of ROS, mitochondria are well-known to constitute a favorite target [74,75].

It was particularly striking that EV cargoes (ferritin and NADPH oxidase), carried away

from the parent cells to protect themselves from oxidative stress, would be responsible for

lysosome alterations in healthy recipient hepatocytes. Precisely, an oxidative reaction, known

to occur in the acidic and hydrolase-enriched lysosome lumen [76], could be triggered by those

cargoes. Thus, degradation of ferritin by lysosome hydrolases could provide LMW iron species

Accepted

27

involved in the catalysis of the Fenton reaction, that converts hydrogen peroxide to the highly

oxidant hydroxyl radical. The latter can then cause lipid peroxidation of lysosome membranes

and consecutive lysosome membrane permeabilization. The use of specific lysosomal iron

chelator and NADPH oxidase inhibitor has led us to show the involvement of EV cargoes, more

precisely ferritin and NADPH oxidase, in PAH-EVs-induced apoptosis of the recipient

hepatocytes. While it is the first time that such a route for iron entry and related damage effects

is proposed, the participation of NADPH oxidase to apoptosis has previously been reported in

endothelial and vascular smooth muscle recipient cells exposed to circulating EVs [56,57].

However, no real explanation of the mechanism involved in the apoptosis induction was

provided by those authors. We now propose that NADPH oxidase carried by EVs could be a

source of superoxide anion and hence hydrogen peroxide to promote, with iron derived from

ferritin, the Fenton reaction needed to induce lysosomal membrane permeabilization (Figure

10).

Finally, iron appeared to have a pivotal role in the process of cell killing by EVs. Indeed,

it can also be found in mitochondria of hepatocytes treated by BP-EVs after translocation from

lysosomes. An iron translocation from lysosomes to mitochondria has already been described

in several types of hepatocyte attack such as ischemia-reperfusion [77], oxidative stress [78] or

drug intoxication [79]), but we show, for the first time, that it can also be related to an EV

endocytosis. In all cases previously reported [77–79], this iron overload of mitochondria was

responsible for mitochondrial ROS formation and permeability transition, and also apoptosis.

It could be confusing to consider EVs as key actors in the progression of PAH-induced

oxidative stress leading to cell death, since hepatocytes exposed to those toxicants release EVs

loaded with pro-oxidant compounds, likely to reduce the intensity of the damage in the parent

cells. Thus, the increase in lipid peroxidation for PAH-exposed hepatocytes was restricted to

only 35 %. However, these EVs were able to induce oxidative stress in healthy untreated

Accepted

28

hepatocytes. To really appreciate the significance of such an effect, it is important to place this

EV release in the particular context of the liver architecture, notably the hepatic lobule, and

PAH exposure. First, the flow of blood entering into the lobules from the portal nodes towards

the central vein through sinusoidal blood vessels lining hepatocytes contributes to a zonation

of the liver. For instance, the pericentral AhR expression [80] and induction of CYP1 enzymes

[81] result in increased exposure of the pericentral hepatocytes to toxic intermediates. The

possible release of EVs by those hepatocytes into the bile flow [82] that circulate in the opposite

direction of the blood flow could thus contribute to the spatial propagation of the PAH toxicity

to healthy hepatocytes located in other zones notably in periportal zone. Second, another issue

worth stressing is that human PAH exposure can be an environmental intermittent exposure,

but EV release due to PAH exposure can propagate the potential hepatotoxicity over a period

of several days even in the absence of PAHs. Therefore, EVs create a vicious cycle, thereby

aggravating oxidative stress by a spatio and temporal way.

5. Conclusion

The present study reports, for the first time, that EVs released after PAH exposure can

mediate lysosome membrane permeabilization of recipient hepatocytes, thereby leading to

oxidative stress and ultimately apoptosis. Considering all the present results, we can suggest

that the EVs released in the extracellular environment by hepatocytes, possibly protecting

themselves from oxidative damage by carrying away deleterious components, could be

transferred with their harmful cargoes to neighboring or even distant healthy hepatocytes and

induce oxidative stress. Thus, EV cargoes, notably iron and NADPH oxidase, appeared as

critical for both EV release and EV-induced oxidative damage. To conclude,

hepatocyte-derived EVs could play a key role in the pathogenesis of liver diseases especially in the field of

toxicant-associated oxidative stress.

Accepted

29 Acknowledgements :

We thank the animal house platform ARCHE (SFR Biosit, Rennes, France) and

Laurence Bernard-Touami for her assistance.

Funding :

This study was financially supported by the "Programme Environnement-Santé-Travail"

of Anses with the funding from ITMO cancer in the context of the Cancer Plan 2014-2019

[EST-2016/1/31], and by the Cancéropôle Grand Ouest/Région Bretagne (CONCERTO

project). Nettie van Meteren is recipient of a doctoral fellowship from the French Ministry for

Higher Education and Research

.

Accepted

30 REFERENCES

[1] M. Yáñez-Mó, P.R.-M. Siljander, Z. Andreu, A. Bedina Zavec, F.E. Borràs, E.I. Buzas, K. Buzas, E. Casal, F. Cappello, J. Carvalho, E. Colás, A. Cordeiro-da Silva, S. Fais, J.M. Falcon-Perez, I.M. Ghobrial, B. Giebel, M. Gimona, M. Graner, I. Gursel, M. Gursel, N.H.H. Heegaard, A. Hendrix, P. Kierulf, K. Kokubun, M. Kosanovic, V. Kralj-Iglic, E.-M. Krämer-Albers, S. Laitinen, C. Lässer, T. Lener, E. Ligeti, A. Linē, G. Lipps, A. Llorente, J. Lötvall, M. Manček-Keber, A. Marcilla, M. Mittelbrunn, I. Nazarenko, E.N.M. Nolte-‘t Hoen, T.A. Nyman, L. O’Driscoll, M. Olivan, C. Oliveira, É. Pállinger, H.A. del Portillo, J. Reventós, M. Rigau, E. Rohde, M. Sammar, F. Sánchez-Madrid, N. Santarém, K. Schallmoser, M. Stampe Ostenfeld, W. Stoorvogel, R. Stukelj, S.G. Van der Grein, M. Helena Vasconcelos, M.H.M. Wauben, O. De Wever, Biological properties of extracellular vesicles and their physiological functions, J. Extracell. Vesicles. 4 (2015) 27066. https://doi.org/10.3402/jev.v4.27066.

[2] M. Simons, G. Raposo, Exosomes – vesicular carriers for intercellular communication, Curr. Opin. Cell Biol. 21 (2009) 575–581. https://doi.org/10.1016/j.ceb.2009.03.007.

[3] G. van Niel, G. D’Angelo, G. Raposo, Shedding light on the cell biology of extracellular vesicles, Nat. Rev. Mol. Cell Biol. 19 (2018) 213–228. https://doi.org/10.1038/nrm.2017.125.

[4] P.B. Devhare, R.B. Ray, Extracellular vesicles: Novel mediator for cell to cell communications in liver pathogenesis, Mol. Aspects Med. 60 (2018) 115–122. https://doi.org/10.1016/j.mam.2017.11.001.

[5] P. Hirsova, S.H. Ibrahim, V.K. Verma, L.A. Morton, V.H. Shah, N.F. LaRusso, G.J. Gores, H. Malhi, Extracellular vesicles in liver pathobiology: Small particles with big impact, Hepatology. 64 (2016) 2219–2233. https://doi.org/10.1002/hep.28814.

[6] K. Sato, L. Kennedy, S. Liangpunsakul, P. Kusumanchi, Z. Yang, F. Meng, S. Glaser, H. Francis, G. Alpini, Intercellular Communication between Hepatic Cells in Liver Diseases, Int. J. Mol. Sci. 20 (2019) 2180. https://doi.org/10.3390/ijms20092180.

[7] I. Garcia-Martinez, N. Santoro, Y. Chen, R. Hoque, X. Ouyang, S. Caprio, M.J. Shlomchik, R.L. Coffman, A. Candia, W.Z. Mehal, Hepatocyte mitochondrial DNA drives nonalcoholic steatohepatitis by activation of TLR9, J. Clin. Invest. 126 (2016) 859–864. https://doi.org/10.1172/JCI83885.

[8] E. Kakazu, A.S. Mauer, M. Yin, H. Malhi, Hepatocytes release ceramide-enriched pro-inflammatory extracellular vesicles in an IRE1α-dependent manner, J. Lipid Res. 57 (2016) 233– 245. https://doi.org/10.1194/jlr.M063412.

[9] C.-Y. Liao, M.J. Song, Y. Gao, A.S. Mauer, A. Revzin, H. Malhi, Hepatocyte-derived lipotoxic extracellular vesicle sphingosine 1-phosphate induces macrophage chemotaxis, Front. Immunol. 9 (2018). https://doi.org/10.3389/fimmu.2018.02980.

Accepted

31

[10] D. Povero, N. Panera, A. Eguchi, C.D. Johnson, B.G. Papouchado, L. de Araujo Horcel, E.M. Pinatel, A. Alisi, V. Nobili, A.E. Feldstein, Lipid-induced hepatocyte-derived extracellular vesicles regulate hepatic stellate cells via microRNA targeting peroxisome proliferator-activated receptor-γ, Cell. Mol. Gastroenterol. Hepatol. 1 (2015) 646–663.

[11] Y.-S. Lee, S.Y. Kim, E. Ko, J.-H. Lee, H.-S. Yi, Y.J. Yoo, J. Je, S.J. Suh, Y.K. Jung, J.H. Kim, Y.S. Seo, H.J. Yim, W.-I. Jeong, J.E. Yeon, S.H. Um, K.S. Byun, Exosomes derived from palmitic acid-treated hepatocytes induce fibrotic activation of hepatic stellate cells, Sci. Rep. 7 (2017). https://doi.org/10.1038/s41598-017-03389-2.

[12] F. Jiang, Q. Chen, W. Wang, Y. Ling, Y. Yan, P. Xia, Hepatocyte-derived extracellular vesicles promote endothelial inflammation and atherogenesis via microRNA-1, J. Hepatol. 72 (2020) 156– 166. https://doi.org/10.1016/j.jhep.2019.09.014.

[13] D. Povero, A. Eguchi, I.R. Niesman, N. Andronikou, X. de Mollerat du Jeu, A. Mulya, M. Berk, M. Lazic, S. Thapaliya, M. Parola, H.H. Patel, A.E. Feldstein, Lipid-induced toxicity stimulates hepatocytes to release angiogenic microparticles that require vanin-1 for uptake by endothelial cells, Sci. Signal. 6 (2013) ra88–ra88. https://doi.org/10.1126/scisignal.2004512.

[14] S. Li, M. Hong, H.-Y. Tan, N. Wang, Y. Feng, Insights into the role and interdependence of oxidative stress and inflammation in liver diseases, Oxid. Med. Cell. Longev. 2016 (2016) 1–21. https://doi.org/10.1155/2016/4234061.

[15] M. Masarone, V. Rosato, M. Dallio, A.G. Gravina, A. Aglitti, C. Loguercio, A. Federico, M. Persico, Role of oxidative stress in pathophysiology of nonalcoholic fatty liver disease, Oxid. Med. Cell. Longev. 2018 (2018) 1–14. https://doi.org/10.1155/2018/9547613.

[16] Z. Chen, R. Tian, Z. She, J. Cai, H. Li, Role of oxidative stress in the pathogenesis of nonalcoholic fatty liver disease, Free Radic. Biol. Med. 152 (2020) 116–141. https://doi.org/10.1016/j.freeradbiomed.2020.02.025.

[17] G. Bodega, M. Alique, L. Puebla, J. Carracedo, R.M. Ramírez, Microvesicles: ROS scavengers and ROS producers, J. Extracell. Vesicles. 8 (2019) 1626654. https://doi.org/10.1080/20013078.2019.1626654.

[18] W. Tan, Y. Zhang, M. Li, X. Zhu, X. Yang, J. Wang, S. Zhang, W. Zhu, J. Cao, H. Yang, L. Zhang, miR-27a-containing exosomes secreted by irradiated skin keratinocytes delayed the migration of unirradiated skin fibroblasts, Int. J. Biol. Sci. 15 (2019) 2240–2255. https://doi.org/10.7150/ijbs.35356.

[19] A. Söderberg, A.M. Barral, M. Söderström, B. Sander, A. Rosén, Redox-signaling transmitted in trans to neighboring cells by melanoma-derived TNF-containing exosomes, Free Radic. Biol. Med. 43 (2007) 90–99. https://doi.org/10.1016/j.freeradbiomed.2007.03.026.

[20] S. Dutta, C. Warshall, C. Bandyopadhyay, D. Dutta, B. Chandran, Interactions between exosomes from breast cancer cells and primary mammary epithelial cells leads to generation of reactive

Accepted

32

oxygen species which induce DNA damage response, stabilization of p53 and autophagy in epithelial cells, PLoS ONE. 9 (2014) e97580. https://doi.org/10.1371/journal.pone.0097580. [21] F.R.M. Stassen, P.H. van Eijck, P.H.M. Savelkoul, E.F.M. Wouters, G.G.U. Rohde, J.J. Briedé,

N.L. Reynaert, T.M. de Kok, B.J. Benedikter, Cell type- and exposure-specific modulation of CD63/CD81-positive and tissue factor-positive extracellular vesicle release in response to respiratory toxicants, Oxid. Med. Cell. Longev. 2019 (2019) 1–9. https://doi.org/10.1155/2019/5204218.

[22] B.J. Benedikter, C. Volgers, P.H. van Eijck, E.F.M. Wouters, P.H.M. Savelkoul, N.L. Reynaert, G.R.M.M. Haenen, G.G.U. Rohde, A.R. Weseler, F.R.M. Stassen, Cigarette smoke extract induced exosome release is mediated by depletion of exofacial thiols and can be inhibited by thiol-antioxidants, Free Radic. Biol. Med. 108 (2017) 334–344. https://doi.org/10.1016/j.freeradbiomed.2017.03.026.

[23] N. van Meteren, D. Lagadic-Gossmann, M. Chevanne, I. Gallais, D. Gobart, A. Burel, S. Bucher, N. Grova, B. Fromenty, B.M.R. Appenzeller, S. Chevance, F. Gauffre, E. Le Ferrec, O. Sergent, Polycyclic aromatic hydrocarbons can trigger hepatocyte release of extracellular vesicles by various mechanisms of action depending on their affinity for the aryl hydrocarbon receptor, Toxicol. Sci. 171 (2019) 443–462. https://doi.org/10.1093/toxsci/kfz157.

[24] S.-H. Cherng, S.-L. Hsu, J.-L. Yang, C.-T.R. Yu, H. Lee, Suppressive effect of 1-nitropyrene on benzo[a]pyrene-induced CYP1A1 protein expression in HepG2 cells, Toxicol. Lett. 161 (2006) 236–243. https://doi.org/10.1016/j.toxlet.2005.10.002.

[25] L. Huc, L. Sparfel, M. Rissel, M.-T. Dimanche-Boitrel, A. Guillouzo, O. Fardel, D. Lagadic-Gossmann, Identification of Na+/H+ exchange as a new target for toxic polycyclic aromatic hydrocarbons, FASEB J. 18 (2004) 344–346.

[26] J.-Q. Ma, C.-M. Liu, Z.-H. Qin, J.-H. Jiang, Y.-Z. Sun, Ganoderma applanatum terpenes protect mouse liver against benzo(α)pyren-induced oxidative stress and inflammation, Environ. Toxicol. Pharmacol. 31 (2011) 460–468. https://doi.org/10.1016/j.etap.2011.02.007.

[27] A. Collin, K. Hardonnière, M. Chevanne, J. Vuillemin, N. Podechard, A. Burel, M.-T. Dimanche-Boitrel, D. Lagadic-Gossmann, O. Sergent, Cooperative interaction of benzo[a]pyrene and ethanol on plasma membrane remodeling is responsible for enhanced oxidative stress and cell death in primary rat hepatocytes, Free Radic. Biol. Med. 72 (2014) 11–22. https://doi.org/10.1016/j.freeradbiomed.2014.03.029.

[28] M. Gorria, X. Tekpli, O. Sergent, L. Huc, F. Gaboriau, M. Rissel, M. Chevanne, M.-T. Dimanche-Boitrel, D. Lagadic-Gossmann, Membrane fluidity changes are associated with benzo[a]pyrene-induced apoptosis in F258 cells: protection by exogenous cholesterol, Ann. N. Y. Acad. Sci. 1090 (2006) 108–112. https://doi.org/10.1196/annals.1378.011.

Accepted